ABSTRACT
Johnson-Laird (1983) specified the general processing assumptions of the mental model theory of reasoning. The mental model theory holds that the basic processes of human reasoning competence are based on the construction and manipulation of mental models: representations of a possible state of affairs as described by structural relationships expressed in the premises. Reasoning by model is broken down into three processing stages (Johnson-Laird, 1983; Johnson-Laird & Byrne, 1991). The processing stages embody the principle of eliminating states of affairs when reasoners try to construct an integrated model of the premises. First, on the basis of the meaning of the premises and other general knowledge triggered during the process of interpretation,
reasoners build a mental model representing a state of affairs. Second, while eliminating inconsistent models, reasoners conjoin the various models of the premises to form an integrated model. The integrated mental model forms the basis to formulate a putative conclusion. Third, in order to validate the putative conclusion, reasoners search for alternative models (possibilities) which might falsify it. If there is such a counterexample, reasoners need to return to the second stage to determine whether there is any conclusion that holds over all the models constructed thus far. When no alternative models can be constructed, the putative conclusion could not be falsified and is valid. The central expectation is that the more models that need to be constructed in order to derive a valid conclusion, the harder the inferential task will be. This prediction gained support from studies of syllogistic reasoning (Johnson-Laird & Bara, 1984), reasoning with multiple qualifiers (Johnson-Laird, Byrne, & Tabossi, 1991), spatial and temporal reasoning (Byrne & Johnson-Laird, 1989; Schaeken, Johnson-Laird, & d’Ydewalle, 1996; Rauh, chap. 9, this volume; Vandierendonck & De Vooght, 1996; Vandierendonck, De Vooght, Desimpelaere, & Dierckx, chap 10, this volume) and prepositional reasoning (Johnson-Laird, Byrne, & Schaeken, 1992).
